AN EXAMINATION OF ABINITIO RESULTS FOR THE HELIUM POTENTIAL-ENERGY CURVE

Obtaining a ground state potential energy curve for helium has been the subject of much research involving empirical, semiempirical, and ab initio methods. In this work, we examine critically recent ab initio potentials proposed for this interaction with respect to their ability to predict certain accurate experimental data. To accomplish this analysis, potentials with a modified HFD-B form were fit to the recent theoretical work of van Duijneveldt and co-workers [Vos, van Lenthe, and van Duijneveldt, J. Chem. Phys. 93, 643 (1990) and Vos, van Mourik, van Lenthe, and van Duijneveldt (to be published)] and Liu and McLean (LM-2) [J. Chem. Phys. 91, 2348 (1989)]. A well depth (epsilon/k = 10.92 K) and a separation at the minimum (r(m) = 2.9702 A) consistent with both determinations were chosen and the properties of helium were calculated based on these potentials. These "mimic" potentials fail to predict the very low temperature He-4 and He-3 virials and one of them [Vos, van Maurik, van Lenthe, and van Duijneveldt (to be published)] also fails to predict the very accurate room temperature viscosity of Vogel [Ber. Bunsenges. Phys. Chem. 88, 997 (1984)]. For a potential which was fit as closely as possible to the LM-2 potential at its maximum suggested depth (epsilon/k = 10.97 K), the virials are satisfactorily predicted. This compromise potential appears to be the best characterization of the He-He interaction in its ability to predict a variety of experimental data as well as being consistent with ab initio results.